U.S. patent number 9,563,852 [Application Number 15/187,963] was granted by the patent office on 2017-02-07 for pest occurrence risk assessment and prediction in neighboring fields, crops and soils using crowd-sourced occurrence data.
This patent grant is currently assigned to ITERIS, INC.. The grantee listed for this patent is ITERIS, INC.. Invention is credited to Dustin C. Balsley, Lori J. Wiles.
United States Patent |
9,563,852 |
Wiles , et al. |
February 7, 2017 |
Pest occurrence risk assessment and prediction in neighboring
fields, crops and soils using crowd-sourced occurrence data
Abstract
A pest and disease modeling framework for precision agriculture
applies weather information, pest biological characteristics, and
crop management data to anonymous crowd-sourced observations of
pest presence for a reporting field. A risk assessment profile of
pest occurrence for targeted fields in proximity to reporting
fields is modeled to generate field-specific measures for pest
management of pest infestation. The pest and disease modeling
framework matches and filters weather and crop information in
infested and pest-free fields based on the anonymous, crowd-sourced
reporting of an existing pest presence, by evaluating similarities
in pest-relevant data. Fields that are similar to infested fields
have the highest risk of infestation, and the modeling framework
provides output data in the form of a prediction of pest occurrence
based on the risk assessment profile.
Inventors: |
Wiles; Lori J. (Fort Collins,
CO), Balsley; Dustin C. (Osage, IA) |
Applicant: |
Name |
City |
State |
Country |
Type |
ITERIS, INC. |
Santa Ana |
CA |
US |
|
|
Assignee: |
ITERIS, INC. (Santa Ana,
CA)
|
Family
ID: |
57908807 |
Appl.
No.: |
15/187,963 |
Filed: |
June 21, 2016 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06Q
10/04 (20130101); A01B 79/005 (20130101); G06N
20/00 (20190101); Y02A 40/10 (20180101) |
Current International
Class: |
G06N
7/00 (20060101); G06N 99/00 (20100101) |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Chen; Alan
Attorney, Agent or Firm: Lazaris IP
Claims
The invention claimed is:
1. A method, comprising: receiving crowd-sourced information
indicating a pest presence in a reporting field and identifying a
pest, in at least one GPS-tagged report from which a GPS receiver
extracts positional coordinates of the reporting field for
correlation of the positional coordinates of the reporting field
with coordinates of one or more targeted fields; obtaining one or
more of pest data that at least includes biological characteristics
of the pest, weather information that at least includes one or more
of recent and current field-level weather data and short-term
weather forecast data for the reporting field and for the one or
more targeted fields, and crop management data that includes at
least one of crop planting dates and crop growth information over
time for both reporting field and the one or more targeted fields,
wherein the crowd-sourced information, the pest data, the weather
information, and the crop management data collectively comprise
input data; analyzing the input data in a plurality of data
processing modules within a computing environment in which the
plurality of data processing modules are executed in conjunction
with at least one specifically-configured processor, the data
processing modules configured to model a field-specific risk of
pest occurrence in the one or more targeted fields, by identifying
a set of descriptors relating to the pest presence in the reporting
field from one or more correlated variable associations of the
input data to develop an expected pest-environment relationship in
the reporting field, and constructing an adaptive and localized
infestation suitability model for the one or more targeted fields
from the set of descriptors to perform a comparison between the
reporting field and the one or more targeted fields from the
expected pest-environment relationship in a unsupervised learning
engine, by 1) selecting one or more environmental and crop
management predictors from the set of the descriptors based on a
multivariate similarity among the correlated variable associations
of the input data, 2) calculating statistical probabilities of
similarity of pest occurrence relative to the selected one or more
environmental and crop management predictors, and 3) continually
updating the modeling of the field-specific risk of pest occurrence
in the one or more targeted fields by identifying the set of
descriptors for the pest-environment relationship, selecting one or
more of the environmental and crop management predictors based on
the multivariate similarity among the correlated variable
associations of the input data, and the statistical probabilities
of similarity of pest occurrence as the crowd-sourced information
is received, to create a risk assessment profile for pest
occurrence for the one or more targeted fields; and generating, as
output data, a pest occurrence prediction from the risk assessment
profile.
2. The method of claim 1, wherein the pest occurrence prediction is
generated for a specified period of time in a current growing
season.
3. The method of claim 2, wherein the specified time period is
determined by the biological characteristics of the pest.
4. The method of claim 1, wherein the pest is an arthropod, a
nematode, a pathogen, or a weed.
5. The method of claim 1, further comprising obtaining soil data
for the reporting field and the one or more targeted fields, the
soil data including one or more of a soil type, pattern of soil
temperature, and pattern of soil moisture content, for at least one
of a current growing season and a period of time extending up to
and including a current growing season.
6. The method of claim 1, wherein the crop management data further
includes plant canopy conditions over time, crop type information,
relative crop maturity, planting date, irrigation application data,
and tillage information.
7. The method of claim 1, further comprising applying the risk
assessment profile to a decision support tool configured to provide
one or more advisories of the risk assessment to a user.
8. The method of claim 1, wherein the one or more advisories
include at least one of an agricultural advisory, a disease
advisory, and a regulatory advisory.
9. The method of claim 1, further comprising updating the pest
occurrence prediction in the one or more targeted fields as
additional crowd-sourced information for additional reporting
fields in a vicinity of a targeted field is received for
constructing the adaptive and localized infestation suitability
model.
10. The method of claim 1, wherein the biological characteristics
of the pest at least include life cycle and development rate of the
pest.
11. A system comprising: a computing environment including at least
one computer-readable storage medium having program instructions
stored therein and a computer processor operable to execute the
program instructions to profile a risk assessment of pest
occurrence within a plurality of data processing modules, the
plurality of data processing modules including: an initialization
module configured to 1) receive crowd-sourced reports that indicate
a pest presence in a reporting field and identify a pest, 2)
correlate positional coordinates of the reporting field and
coordinates of one or more targeted fields from at least one
GPS-tagged report in the crowd-sourced information from which a GPS
receiver extracts the positional coordinates of the reporting
field, and 3) identify and obtain additional input data relative to
the pest presence, including one or more of pest data that at least
includes biological characteristics of the pest, weather
information that at least includes one or more of recent and
current field-level weather data and short-term weather forecast
data for both the reporting field and for the one or more targeted
fields, and crop management data that includes at least one of crop
planting dates and crop growth information over time for both
reporting field and the one or more targeted fields; one or more
modules configured to identify a set of descriptors relating to the
pest presence in the reporting field from one or more correlated
variable associations of the crowd-sourced observations and the
additional input data to develop an expected pest-environment
relationship in the reporting field, and construct an adaptive and
localized infestation suitability model for the one or more
targeted fields from the set of variables to perform a comparison
between the reporting field and the one or more targeted fields
from the expected pest-environment relationship, in an unsupervised
learning model that 1) selects one or more environmental and crop
management predictors from the set of descriptors based on a
multivariate similarity among the correlated variable associations
of the input data, 2) calculates statistical probabilities of
similarity of pest occurrence relative to the one or more
environmental and crop management predictors, and 3) continually
updating the set of descriptors for the pest-environment
relationship, the selection of the one or more of the environmental
and crop management predictors based on the multivariate similarity
among the correlated variable associations of the input data, and
the statistical probabilities of similarity of pest occurrence as
the crowd-sourced reports are received, and create a risk
assessment profile for pest occurrence for the one or more targeted
fields; and a pest occurrence prediction module configured to
predict a pest occurrence from the risk assessment profile.
12. The system of claim 11, wherein the risk assessment profile is
applied to an agricultural support tool to generate one or more
advisories for the targeted field.
13. The system of claim 12, wherein the one or more advisories
include at least one of an agricultural advisory, a disease
advisory, and a regulatory advisory.
14. The system of claim 11, wherein the pest occurrence prediction
is generated for a specified period of time in a current growing
season.
15. The system of claim 11, wherein the specified time period is
determined by the biological characteristics of the pest.
16. The system of claim 11, wherein the pest is an arthropod, a
nematode, a pathogen, or a weed.
17. The system of claim 11, wherein the additional input data
further includes soil data for the reporting field and the one or
more targeted fields, the soil data including one or more of a soil
type, a pattern of soil temperature, and a pattern of soil moisture
content, for at least one of a current growing season and a period
of time extending up to and including a current growing season.
18. The system of claim 17, wherein the crop management data
further includes plant canopy conditions over time, crop type
information, relative crop maturity, planting date irrigation
application data, and tillage information.
19. The system of claim 11, wherein the risk assessment profile is
updated for the one or more targeted fields as additional
crowd-sourced information for additional reporting fields in a
vicinity of a targeted field is received for constructing the
adaptive and localized infestation suitability model.
20. The method of claim 11, wherein the biological characteristics
of the pest at least include life cycle and development rate of the
pest.
21. A method of assessing a field-specific risk of pest occurrence
in a targeted field, comprising: identifying 1) a pest presence in
a reporting field, and 2) and positional coordinates of the
reporting field, from crowd-sourced information that includes at
least one GPS-tagged report from which a GPS receiver extracts the
positional coordinates of the reporting field for correlation of
the positional coordinates of the reporting field with coordinates
of one or more targeted fields; defining an expected
pest-environment relationship for the reporting field by
identifying a set of descriptors relating to the pest presence in
the reporting field from one or more correlated variable
associations of the crowd-sourced information, and one or more of
pest data that at least includes biological characteristics of a
pest identified in the reporting field, weather information that at
least includes one or more of recent and current field-level
weather data and short-term weather forecast data for the reporting
field, and crop management data that includes at least one of crop
planting dates and crop growth information over time for the
reporting field; matching the expected pest-environment
relationship for the reporting field at least with weather
information that at least includes one or more of recent and
current field-level weather data and short-term weather forecast
data for the one or more targeted fields, and crop management data
that at least one of crop planting dates and crop growth
information over time for the one or more targeted fields, by
constructing an adaptive and localized infestation suitability
model for the one or more targeted fields for performing a
comparison between the reporting field and the one or more targeted
fields within an unsupervised learning engine, configured to 1)
select one or more environmental and crop management predictors
from the set of descriptors based on a multivariate similarity
among the correlated variable associations, 2) calculate
statistical probabilities of similarity of a pest occurrence
relative to the selected one or more environmental and crop
management predictors, and 3) continually updating the set of
descriptors for the pest-environment relationship, selection of the
one or more environmental and crop management predictors, and the
statistical probabilities of similarity of a pest occurrence, as
either of a pest presence in a reporting field or positional
coordinates of the reporting field are identified from
crowd-sourced information to profile a risk assessment of pest
occurrence in the one or more targeted fields; and estimating a
probability of pest occurrence for the one or more targeted fields
from the profile of risk assessment.
22. The method of claim 21, wherein the probability of pest
occurrence is estimated for a specified period of time in a current
growing season.
23. The method of claim 22, wherein the specified time period is
determined by the biological characteristics of the pest.
24. The method of claim 21, wherein the pest is an arthropod, a
nematode, a pathogen, or a weed.
25. The method of claim 21, further comprising identifying and
obtaining additional input data that includes soil data for the
reporting field and the one or more targeted fields, the soil data
including one or more of a soil type, a pattern of soil
temperature, and a pattern of soil moisture content, for at least
one of a current growing season and a period of time extending up
to and including a current growing season.
26. The method of claim 21, wherein the crop management data
further includes plant canopy conditions over time, crop type
information, relative crop maturity, planting date, irrigation
application data, and tillage information.
27. The method of claim 21, further comprising applying the profile
to a decision support tool configured to provide one or more
advisories of the risk assessment to a user.
28. The method of claim 21, wherein the one or more advisories
include at least one of an agricultural advisory, a disease
advisory, and a regulatory advisory.
29. The method of claim 21, further comprising updating the
probability in the one or more targeted fields as additional
crowd-sourced information for additional reporting fields in a
vicinity of a targeted field is received for constructing the
adaptive and localized infestation suitability model.
30. The method of claim 21, wherein the biological characteristics
of the pest at least include life cycle and development rate of the
pest.
Description
FIELD OF THE INVENTION
The present invention relates to assessing and predicting an
occurrence of pests in crops, soils, and fields. Specifically, the
present invention relates to a system and method of modeling a risk
of pest occurrence based on an interaction of weather information
and agricultural information for a targeted field, and known pest
presence data in neighboring fields.
BACKGROUND OF THE INVENTION
Pests such as weeds, insects, and pathogens occurring in fields,
crops and soils are constant problems for the agricultural
industry. The presence of a pest in a field is a result of weather
interacting with crop and soil management activity. Yet there are
few weather-driven pest or disease models available for management
decisions in relation to major crops. Also, field-specific risk
assessments are only available for very few pests because the
required models of pest biology are expensive to develop, despite
the high value in crop output. Other issues include a lack of data
for proof of concept to validate existing models or to design new
models, and inherent apprehension among growers and landowners in
admitting pest infestation in their fields and crops.
Some models have been developed to predict pest infestation based
on weather variables. For example, development of insect or an
insect population has been modeled with growing degree days to
predict when the insect or a substantial portion of the insect's
population will be present in a field. Disease has been predicted
by comparing current or recent climatic factors such as temperature
and leaf wetness over a few hours or days against measured climatic
factors that are known to be favorable for a particular pathogen.
These models, however, do not take into account forecasts of
weather data and predictions, and also ignore other factors that
interact with weather and lead to infestation. For example, crop
management in particular may alter the impact of weather. Also, the
crop is often susceptible at certain growth stages and crop
development, like insect development, is influenced by variances in
weather conditions.
Existing approaches are limited at least in part because attempting
to quantify and model all these interactions is overwhelming. Even
attempting to identify which are the most important factors that
should be modeled is a major challenge due to the
constantly-changing parameters during a growing season. Moreover,
changing crop management, like different tillage methods or
planting of seed varieties across different seasons, requires new
studies and constant updating of models.
One existing approach to pest management is simply to compare
climatic factors in un-infested and infested fields. Climatic
factors in such an approach are explicitly handled as time series
of data, and compared using methods to assess the similarity of
time series. For example, a similarity of the pattern of average
temperature over a series of days in infested versus un-infested
fields would be calculated. Growing degrees days is one measure of
temperature over a period, but growing degree days but does not
explicitly consider such a pattern. The same value of growing
degree days may be accumulated from several different patterns of
weather over a set of days.
Other existing approaches to providing information regarding an
agricultural pest infestation include websites that map
observations of pest scouting or counts of pests from traps, along
with weekly emails describing pest problems reported to extension
specialists. However, constraints of cost and data privacy mean
universities cannot provide information about crop management
associated with specific pest-infested fields. Managers must
therefore guess at the relevance of the information to their own
fields, greatly reducing the reliability of such information.
Also, weather and crop information is also often imprecise on such
websites, and therefore mapping only relates pest occurrence in a
general way to weather. The resulting resolution of the information
is poor, and pest presence is not related to specific weather
variables or specific crop management activities. The manager must
therefore speculate how crop management and growth stage factors in
the mapped infested areas mitigate or promote pest presence, and
for the number of fields in which the pest was observed. Managers
must also speculate as to what weather conditions promoted the
presence of the pest, and whether there are similar conditions
forecast or prevailing in his or her managed fields. Managers will
therefore not be able to accurately determine what crop management
activities mitigate or promote pest infestation, nor will they be
able to determine which of fields have a high priority for scouting
or treatment.
Many techniques are available for obtaining crowd-sourced
information. Often, crowd-sourced information is crucial to
containing a spread of a pest or disease over a wider area because
of the real-time nature of such ground truth observations. However,
the effectiveness of crowd-sourcing for pest management is sharply
limited by the reluctance of growers to reveal the presence of
pests in their fields to others. This reluctance is harmful, as
speedy knowledge of an infestation in a nearby field can help the
wider region contain the spread and avoid costly damage with quick
action. There is no existing approach to pest management that
leverages crowd-sourced reporting anonymously, so that growers can
feel comfortable with accurate reporting of pest and disease issues
in their own crops and fields.
BRIEF SUMMARY OF THE INVENTION
It is therefore one objective of the present invention to provide a
system and method of modeling a risk of pest occurrence in a
targeted field. It is another objective of the present invention to
provide a system and method of predicting pest occurrence in a
targeted field. It is a further objective to assess a risk of, and
predict, a pest occurrence in a targeted field based on
crowd-sourced information of pest presence in neighboring or nearby
fields. It is yet another objective to provide a system and method
of examining an interaction of weather conditions and crop
management practices to model a risk of pest occurrence in a
targeted field over a specific time period, and generate a
prediction of the pest occurrence in the targeted field. It is
still another objective to provide an indication to growers,
landowners, crop advisors, and other responsible entities of a
possible pest presence in a targeted field to enable one or more
responsive management actions. It is yet another objective of the
present invention to provide an advisory service with recommended
management actions and other alerts and notifications to such
growers, landowners, crop advisors and other responsible entities
where this is a risk or prediction of pest presence in a targeted
field.
The present invention applies precise weather data and
field-specific information about crop management to anonymous
crowd-sourced observations of infested fields reporting a pest
presence, and models that information for pest management in
un-infested, or targeted fields. The present invention also
provides a crowd-sourced pest and disease analytical tool that
generates a risk assessment profile of pest occurrence and a
prediction of pest occurrence for targeted fields, and generates
field-specific measures for pest management of pest occurrence.
The present invention matches weather information and field
information in infested and pest-free fields based on these
anonymous crowd-sourced observations by evaluating similarities in
pest-relevant weather data and crop management data in infested and
pest-free fields. Similarities may also be evaluated in other types
of information, for example in soil conditions, plant canopy
temperature and moisture data, and simulated growth stage. The
present invention incorporates analytical tools to continually
update calculations in this evaluation of similarities to assess
the risk of pest occurrence for targeted fields as pests are
observed in more reporting fields.
The crowd-sourced pest and disease analytical tool and model
receives a crowd-sourced report of a pest presence in a field, and
accesses field-specific weather data and crop management
information for each infested reporting field. The present
invention also accesses pest biology data for each reported pest.
Based on such knowledge, the present invention filters the weather
and crop management information to identify a set of descriptors
and relevant variables thereof, and select one or more predictors,
from correlated variable associations of the input data impacting a
pest-environment relationship for that field. The present invention
then calculates the similarity of reporting and targeted fields for
specific pests to profile a risk assessment for targeted fields.
Fields that are similar to infested fields have the highest risk of
infestation, and the present invention provides output data in the
form of a prediction of pest occurrence based on the risk
assessment profile.
The present invention develops an infestation suitability model
that is initiated by selecting, from all available crop management
and weather data about infested fields, that data which is
estimated to provide appropriate correlations with pest presence.
This may be thought of as an a priori selection of potential
descriptors, based on knowledge of the population and spatial
dynamics of the pests. The present invention then puts the selected
descriptors in an unsupervised learning method engine (or, an
ensemble of such methods) to look for patterns in the selected data
and the relation to characteristics of targeted fields to develop
one or more environmental and crop management predictors based on a
multivariate similarity of variable values among the selected set
of descriptors. This narrows the set of descriptors and determines
their relative importance, and in some cases, the form of the
relationship between the environmental variable and the likelihood
of a pest problem. This resulting infestation suitability model is
used to develop a risk assessment profile, which is applied to
perform a calculation of the risk. The risk assessment profile may
also be applied to generate a ranking of risk of targeted fields.
Every time additional observations of pest presence are received,
the present invention enhances its predictive capabilities by
modifying the pest-environmental relationship described by the
infestation suitability model, so that the model is both adaptive
and dynamic.
In an alternative embodiment, observations of both a pest presence
in, and a pest absence from, a reporting field may be used to model
a likelihood of a pest problem. In such an embodiment, feedback
from users on whether a prediction was correct (i.e., did the pest
infestation occur) provides presence (or, absence data) for a
post-ante analysis to refine predictor selection as to descriptors
comprised of field variables and their relative importance. Such
feedback may also aid in refining a definition of an
agro-ecological zone, and provide information for selection of the
most appropriate modeling methods.
Other objects, embodiments, features and advantages of the present
invention will become apparent from the following description of
the embodiments, taken together with the accompanying drawings,
which illustrate, by way of example, the principles of the
invention.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
The accompanying drawings, which are incorporated in and constitute
a part of this specification, illustrate several embodiments of the
invention and together with the description, serve to explain the
principles of the invention.
FIG. 1 is a block diagram illustrating system architecture
components in a crowd-sourced pest and disease model according to
one embodiment of the present invention; and
FIG. 2 is a flowchart of steps in a process of performing a
crowd-sourced pest and disease model according to another
embodiment of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
In the following description of the present invention reference is
made to the exemplary embodiments illustrating the principles of
the present invention and how it is practiced. Other embodiments
will be utilized to practice the present invention and structural
and functional changes will be made thereto without departing from
the scope of the present invention.
FIG. 1 is a system architecture diagram for a crowd-sourced pest
and disease model 100 for predicting a presence of, and profiling a
risk assessment for, a pest or disease 102 in a particular, or
targeted field 106, according to one embodiment of the present
invention. The crowd-sourced pest and disease model 100 is
performed within one or more systems and/or methods that includes
several components, each of which define distinct activities
required to apply real-time, field-level observations
representative of a pest presence in a reporting field 104,
localized weather conditions, together with long-range
climatological and/or meteorological forecasting, crop management
data, and pest biology data to analyze and assess a risk of
infestation. This risk analysis is used to generate, in one
embodiment, a prediction of pest occurrence in targeted field(s)
106 and provide diagnostic support for pest management as well as
farm and harvest operations.
The risk of pest occurrence for individual fields is estimated in
the present invention by the similarity of un-infested fields (or
targeted fields 106) with fields that have been reported, using
anonymous crowd-sourced information, as infested locally and during
the current season (or reporting fields 104). For each pest 102,
the present invention develops variables related to weather
information, crop management actions, and other field, crop and
soil characteristics that are associated with the presence of the
pest 102. Pest biology data may be used to filter these variables,
and statistical measures of similarity between un-infested fields
and infested fields are calculated for each variable, based on
weighted vectors of predictive variables that influence pest
infestation. The risk of occurrence in a field is then modeled from
these measures of similarity for a targeted field 106. A profile of
risk assessment is developed, which may include an estimated risk
category of pest infection for each field and/or a ranking of
multiple fields by risk. The crowd-sourced pest and disease model
100 therefore combines anonymous crowd-sourcing of pest presence
and related information such as field-specific weather data, and
models this information to create localized, dynamic measures of
pest risk for individual fields. The crowd-sourced information of
reported infestations is anonymous, so that users do not know which
fields in their vicinity have reported a pest presence. It is to be
understood that the word "field" may also include an area, rather
than simply a specific field with defined boundaries. Therefore, in
one aspect of the present invention (and by way of example), one or
more targeted fields 106 may comprise an arbitrarily-sized area.
Similarly, a reporting field may likewise comprise a reporting
area.
The crowd-sourced pest and disease model 100 performs these
functions by ingesting, retrieving, requesting, receiving,
acquiring or otherwise obtaining input data 110 that includes
reported pest presence data 111 for one or more reporting fields
104. Pest presence data 111 indicates that a specific pest 102 is
present in the reporting field 104. Such data 111 is, as noted
herein, provided in one or more anonymous crowd-sourced
observations, and is reported electronically in a variety of
different ways. For example, anonymous crowd-sourced observations
may be reported using applications resident on or accessed from
computing devices, or via one or social media tools. The pest
presence data 111 may be reported directly by growers, land owners,
crop advisors, or other responsible entities for the reporting
field 104, or automatically. Pest presence data 111 may also
include abundance data indicative of a pest density or severity
rating when a pest is present. Regardless, it is to be understood
that the present invention enables pest presence data 111 to be
reported anonymously, so that recipients of output data 120
representing a prediction of pest infestation 122 for the targeted
field(s) 106 are unaware of either the source of reported pest
presence data 111 or the specific reporting field 104.
Crowd-sourced observations in pest presence data 111 may take many
forms, and may be reported in multiple ways. For example, users may
input indications or reports of a pest presence via an application
or other electronic interface (or other method, as described
further herein). Users may be able to type in their reports, or
speak what they want to indicate, into a mobile or other computing
device. Users may further be able to communicate their observations
of a pest presence via text messaging, a telephone call, and any
other means of communication. Such an application or interface may
include one or more pull-down menus from which users can select a
type of pest and other information that is used to compile a report
of a pest 102 for pest presence data 111. Users may upload a
previously-written report or other document, such as a spreadsheet,
that contains information used to compile a report or otherwise
indicate a pest 102 is present in a field 104. Users may also be
able to upload a photo or a video of a pest 102, and it is to be
understood that the present invention is capable of identifying the
pest 102 from such a photo (or document, spreadsheet, typed or
spoken language, etc.) to produce pest presence data 111. Social
media feeds may also be used as crowd-sourced observations of a
presence of a pest 102. Crowd-sourced observations may therefore be
either manual, automated or automatic, or both. The present
invention is therefore not to be limited to any one type of
crowd-sourced observation, or to any one way or means of inputting
reports or indications of a pest 102 in a field 104.
It should be noted that the word pest, as used herein, refers to
many different types of nuisances affecting crops, plants, soils,
and fields. Therefore, pests 102 may include, but not limited to,
arthropods (for example, insects), nematodes, weeds, or pathogens
(such as for example bacteria, fungi, and viruses). Unless
otherwise indicated herein, pest 102 shall therefore refer
collectively to all of these nuisance types.
Input data 110 also includes field data 112 for both reporting
fields 104 and targeted fields 106. Field data 112 includes various
field characteristics, such as field, soil and/or crop-related
management actions taken. For example, field data 112 may include
historical or recent tillage practice, such as the timing and type
of tillage employed and equipment used. Treatments applied to the
field may also be included in the field data 112 (such as
nutrients, biologicals, or chemicals, and the timing and type of
application), as well as a history of crops and seeds planted in
prior growing seasons, and any prior pest or disease infestations.
Field data 112 may further include water-related information such
as field drainage characteristics, groundwater, watershed and
aquifer data, and information on prior and recent irrigation
practice. Field data 112 may further include whether a field is
managed as reduced or no-tillage, other crop analytics such as the
crop variety and susceptibility to a specific pest, row spacing,
coverage/population, and the type of equipment used in the field,
and the management practices attendant to such equipment.
Field characteristics may also include landscape information such
as an identification of vegetation in areas adjacent to a planted
crop, and soils information for the reporting fields 104 and
targeted fields 106, as well as cropping history and pest presence
in the fields and surrounding landscape in previous seasons. Field
characteristics may further include information regarding land
adjacent to reporting and/or targeted fields, such as for example
types of trees/vegetation, water sources such as streams, and types
of crops planted or growing nearby. Regardless, this may include
information for both a current and past growing seasons, as well as
soils information for fields between seasons, available for example
from public database collections or any other source of such
data.
Input data 110 may also include GPS data 113 that enables the
crowd-sourced pest and disease model 100 to correlate reporting
fields 104 and targeted fields 106. Such GPS data 113 enables GPS
receivers to determine positional coordinates and/or boundaries of
both reporting fields 104 and targeted fields 106 and their
proximity to each other. This allows the crowd-sourced pest and
disease model 100 to determine a geographical correlation for
profiling the risk assessment 145 and prediction of pest occurrence
122 in targeted fields 106, based on the reported pest presence
data 111 in reporting fields 104, as discussed further herein.
Other methods of correlating fields 104 and 106 may also be
utilized and are within the scope of the present invention.
In one embodiment of the present invention, all reported pest
presence data 111 is provided via an application or other
electronic interface (or other method, as described further
herein), and a requirement to access the anonymous crowd-sourced
observations comprising the reporting pest presence data 111 for
modeling a pest occurrence is that the user wishing to know a
possibility of pest occurrence must provide a set of information
about one's field 106, such as for example crop type, seed variety,
planting date, soil conditions, and tillage or other crop
management characteristics, etc. The user may also be required to
report any prior or existing pest infestations in other fields
owned or managed. Users of the present invention, which may be
packaged as a subscription-based software-as-a-service, may also be
required to designate positional coordinates of their fields (the
targeted fields 106). In this manner, the present invention may
continually collect input data 110 for modeling within the various
data processing components 132.
Input data 110 also comprises meteorological and climatological
data 114, which at least includes recent and current field-level
weather data and short-term weather forecast data for both
reporting fields 104 and targeted fields 106, and may further
include historical, predicted, and other weather information, from
many different sources as noted further herein. Recent and current
field-level weather data may represent in-situ or remotely-sensed
observations for one or more of the reporting fields 104 and
targeted fields 106, and may itself be derived from or provided
directly via one or more crowd-sourced observations. The
meteorological and climatological data 114 may be ingested into the
present invention in raw, unprocessed form, or as processed data in
the form of modeled, predicted or forecasted weather data over
particular periods of time, such as short-range weather predictions
and long-range, extended, and/or dynamical weather forecasts.
It is to be noted that the meteorological and climatological data
114 may include different data for both of the reporting fields 104
and the targeted fields 106. Additionally, historical weather
information may include data for at least one of the previous
season, the period between the previous season and the current
season, and the current season. Current field-level weather data
for reporting fields 104, and nearby fields, may include both
historical weather and short-term weather forecasts of field-level
weather for both of the targeted fields 106 and fields nearby the
reporting fields 104. Furthermore, meteorological and
climatological data 114 may include regional weather data that is
specifically relative to pests 102 that do not winter in the
reporting fields 104 and targeted fields 106, such as those pests
that migrate seasonally.
Input data 110 also includes pest and disease data 115. This type
of data is important for an understanding of factors influencing
whether or not an infestation in a particular field will occur.
Pest and disease data 115 may include pest biology and other
characteristics, such as for example life cycle and development
rate. Other biological characteristics may include habitat, range,
reproduction rate, breeding habits, and phenotypic plasticity. Such
other biological characteristics may further include other features
of the pest life cycle, such as for example does the pest
overwinter in the soil, does the pest blow in from the south, or
come from other hosts. Regardless, these other biological
characteristics may also serve to assist in selection of weather
and crop management information for predictors of whether a pest
occurrence will occur in the one or more targeted fields 106.
Crop-related pest biological characteristics include feeding
traits, plant injuries inflicted, and preferred host environments.
Other biological characteristics include tolerance to temperature
and moisture conditions, and resistance to chemical applicants such
as insecticides or herbicides. Other factors may also play a role
in influencing a pest infestation, such as the presence of
predators in a particular field. All of these may also influence
the development and selection of environmental and crop management
predictors 160 in the initialization of an infestation suitability
model 148. Biological characteristics in pest and disease data 115
may further include knowledge of spatial dynamics of the pest life
cycle, and environmental constraints on pest infestation. This may
further include knowledge of crop management practices or weather
patterns correlated with pest infestation.
Input data 110 may further include crop and planting data 116, such
as crop type, seed type, planting data, growing season data and
projections, projected harvest date, crop canopy and soil
conditions over time, relative maturity, planting or emergence
date, crop temperature, crop moisture, seed moisture, plant depth,
row width, and any other crop and plant information that may be
used to profile the risk assessment 145, and the pest occurrence
predictions 122 forming the output data 120. Crop and planting data
116 may further comprise crop management information that
incorporates all of the above types of data. Regardless, crop and
planting data 116 may be provided as output data from one or more
of phenology models of crop and plant growth, and other methods of
predicting crop and plant growth over the course of a growing
season, such as continual crop development profiling of the like
disclosed in U.S. Pat. No. 9,131,644. Similarly, harvest data may
be provided as output data from one or more models of
harvestability, such as those disclosed in U.S. Pat. No. 9,076,118.
Crop and planting data 116 may also provide further information
about crop management actions for the reporting fields 104 and
targeted fields 106.
Additional input data 110 may include soil data 117, such as for
example the soil type, soil temperature, soil moisture content,
soil porosity, soil pH, soil profile, and mineral content, such as
for example its sodicity. Soil data 117 may include a temporal
component, so as to represent conditions in a current growing
season, or for a period of time extending until a following growing
season, or for a prior growing season(s). Soil data 117 may be
imported from one or more external database collections, such as
for example the USDA NRCS Soil Survey Geographic (SSURGO) dataset
containing background soil information as collected by the National
Cooperative Soil Survey, or from one or more models configured to
profile soil structure and composition. Soil data 117 may also be
provided from growers or landowners themselves (or other
responsible entities), from soil advisory tools, from farm
equipment operating in a field, and any other source of such
information.
Crop and planting data 116, biological characteristics for a pest
102, and other input data 110 may further define a specific time
period for the prediction 122 of pest occurrence in a targeted
field 106. For example, the present invention may generate a
prediction 122 that a pest occurrence will materialize within x
number of days, as a result of the growth stage of a crop in the
targeted field 106, the known life cycle of the pest 102, and known
crop management activity. The present invention may therefore apply
the risk assessment profile 145 to identify a specific time period
for an infestation, in addition to a prediction 122 that an
infestation will occur.
Input data 110 may also include many other types of data that can
be used by the crowd-sourced pest and disease model 100 to profile
the risk assessment 145 and generate the pest occurrence prediction
122 in targeted fields 106. Such other data may include imagery
data 118, such as remotely-sensed satellite imagery data and
remotely-captured drone imagery data from orbiting satellites or
remotely powered vehicles. When processed, such imagery data 118
may provide details at a field-level resolution, which can be used
by the crowd-sourced pest and disease model 100 to improve both the
risk assessment profile 145 and pest occurrence prediction 122 in
the present invention.
The input data 110 may further include background data for the
steps involved in identifying descriptors 161 and selecting
predictors 160 from multivariate similarly 149 of variables thereof
for unsupervised learning, as discussed further herein. Such
background data may be acquired from users registering and
reporting on all fields that they have. This background data is
also anonymous and based on crowd-sourcing. The present invention
uses the information on reported fields without the pest, but with
reasonable conditions to assume that the pest might be present, as
the background data. Reasonable conditions are defined based on
what is known about the pest. For example, the present invention
may be configured to not use a field as background information if
the crop has not reached the susceptible growth stage, or if the
model, using weather and the planting date, predicts the crop is
not at the susceptible growth stage.
The positional coordinates of reporting fields 104 and targeted
fields 106 may identify a specific agro-ecological zone for a
localized modeling of infestation suitability. While it is to be
expected that weather patterns and crop management are similar
within a common agro-ecological zone, it should be noted that an
agro-ecological zone may be defined by either or both environment
and management practices, rather than merely using distance alone.
Therefore, GPS coordinates may serve as predictors 160 in the
infestation suitability model 148 depending, for example, upon how
such as agro-ecological zone is defined.
Regardless of the type input data 110, information ingested into
the present invention may include, in addition to anonymous
crowd-sourced observations, one or more of image-based data,
vehicular data, sensor data, and information from other third party
systems. Image-based data may be derived from systems such as video
cameras, and vehicular data may be generated from one or more
vehicle-based sensing systems, including those coupled to computing
systems configured on farm equipment, or those systems configured
to gather weather data from mobile devices present within vehicles,
such as with mobile telephony devices and tablet computers. Sensor
data may be provided from one or more sensors in or near a field,
such as sensors configured on farm equipment or positioned
throughout a field. Crowd-sourced observations, as noted above, may
be provided from multiple sources, such as for example growers,
farmers, land owners, equipment operators, crop advisors or
consultants, and any other responsible entities, and such data may
be provided via oral or written reports, or electronically using
mobile telephony devices or tablet computers, or any other
computing devices that incorporate software tools such as mobile
applications for accessing and using social media feeds. Regardless
of the method of collection or source, however, it is to be
understood that anonymity is maintained for reporting users by
never showing where the pest was observed to users wishing to know
a risk or prediction of infestation for targeted fields 106.
The input data 110 is applied to a plurality of data processing
modules 132 within a computing environment 130 that also includes
one or more processors 134 and a plurality of software and hardware
components. The one or more processors 134 and plurality of
software and hardware components are configured to execute program
instructions or routines to perform the functions of the
crowd-sourced pest and disease model 100 described herein, and
embodied by the plurality of data processing modules 132.
The plurality of data processing modules 132 in computing
environment 130 include a reporting component 140 which processes
incoming pest presence data 111 and determines whether to proceed
with performing the steps of the crowd-sourced pest and disease
model 100. This may be accomplished by analyzing the positional
coordinates of reporting fields 104 and targeted fields 106, for
example where reporting data 111 in the crowd-sourced observations
includes GPS-tagged occurrence data. GPS data 113 may also be
requested by the present invention, for example via data ingest
component 141.
The plurality of data processing components 132 may therefore also
include such a data ingest component 141, which is configured to
perform ingest, retrieval, request, reception, acquisition or
obtaining of input data 110. The plurality of data processing
modules 132 may also include a pest-environment identification
component 142, which processes the input data 110 to identify one
or more predictive field variables for the set of descriptors 161,
of an expected pest-environment relationship 143. This is performed
by examining associations between the types of input data 110 to
identify factors explaining the conditions leading to the pest
occurrence in the reporting field 104. When crowd-sourced
information that includes pest presence data 111 is received into
the present invention, the crowd-sourced pest and disease model 100
may pre-select variables for these descriptors 161 based on known
characteristics of the pest 102 or the reporting field 104.
The data ingest component 141 may be configured to perform a number
of different functions. For example, the data ingest component 141
may be configured to convert raw information from reporting fields
104 and targeted fields 106 into descriptors 161 and the predictors
160 for the infestation suitability model 148, and initialize
algorithms for building the infestation suitability model 148. It
may also be configured to monitor incoming reports of pest
presence, and appropriately trigger rebuilding of descriptors 161
and predictors 160 for the infestation suitability model 148 and
generation of a risk assessment profile 145. The data ingest
component 141 may also monitor incoming requests to generate risk
assessment profiles for the one or more targeted fields 106.
The present invention provides, in one embodiment thereof, a
generalized approach to pest and disease modeling in precision
agriculture, in which prediction of a likelihood of infestation is
determined from similarities between a targeted field 106 and
infested, reporting fields 104 based on weather, crop, field and
landscape characteristics correlated with a presence or abundance
of the pest in those infested fields 104.
This generalized approach includes several steps in evaluating the
likelihood of infestation. One such step is to identify the
relative correlation of pest infestation from variables relative to
weather, field, soil, crop, landscape, and other characteristics of
infested fields 104. It is contemplated that such variables may be
pre-selected based on known characteristics of the pest 102 or the
reporting field 104, such as those previously learned or identified
associations among the input data 110. Regardless, correlated
associations of these variables among the input data 110 are used
to identify a set of descriptors 161 to develop an
observed/expected pest-environment relationship in the infested
fields 104. Another step is to characterize the observed
relationship between pest infestation and the correlated variable
associations. The observed relationship may represent, for example,
an average value of a variable, or identifying a representative
response.
A further step assesses the similarity of variables of a targeted
field 106 to infested fields 104 based on the relative correlation
of variable associations and observed relationship. This is an
assessment of multivariate similarity 149 of variable values, and
may apply many different statistical measures of similarity, such
as distance or correlation. The appropriate measure depends on the
type of variable (i.e. continuous, categorical or binary).
The present invention applies unsupervised machine learning in
implementing each of these steps, and in many of these approaches,
these steps are not distinct. Regardless, some form of unsupervised
machine learning is applied to the descriptors 161 representing
variables of the infested fields 104, and the assessment of
multivariate similarity 149 of variables of a targeted field 106 to
those of the infested fields 104. One way similarity may be
measured is by comparing the relevant variables of targeted fields
106 to a summary of the response for the infested fields 104.
Another exemplary approach compares the value for the targeted
field 106 to all values for infested fields 104 with an algorithm
based on self-organized maps. A further example of accomplishing
these steps involves feature selection and weighting, by
characterizing the relationship of a relevant characteristic with
an average value and measuring the similarity of a targeted field
106 and the average value for infested fields 104 for each
individual characteristic. An aggregate similarity measure is
calculated using the vector of weights from feature selection and
similarity values of individual characteristics.
Still another example of unsupervised machine learning to perform
these steps uses clustering. Suitable clustering methods include
knn clustering, spectral clustering or neural network clustering.
The prediction of risk of pest infestation is based on which the
number of infested fields 104 in each cluster and the cluster
membership of a targeted field 106 when it is evaluated by the
clustering method used with the infested fields.
The crowd-sourced pest and disease model 100 therefore also
includes a unsupervised learning engine 144, which is configured to
create a risk assessment profile 145 by comparing characteristics
of reporting fields 104 and targeted fields 106 in environment and
crop management predictors 160 selected from descriptors 161 of
field variables identified in the pest-environment relationship
143. The risk assessment profile 145 can be generated using many
different methods, as noted in detail below. Regardless of the
specific method, the unsupervised learning engine 144 may apply one
or more of statistical analyses 146 and other mathematical
processes 147 to create an infestation suitability model 148 from
the pest presence data 111. The infestation suitability model 148
may be considered as application of artificial intelligence, for
example in one or more models 165 that automatically and
continually identify additional descriptors 161 and select
additional environmental and crop management predictors 160, as
well as any coefficients for those predictors 160 related to the
multivariate similarity analysis 149, for the pest-environment
relationship 143 as additional crowd-sourced information is
received.
One such method involves assigning weighted vectors of the field
variables, where weights denote the importance of each variable
identified in the pest-environment relationship. The unsupervised
learning engine 144 models these weighted vectors of field
variables by performing statistical analyses 146 and other
mathematical processes 147 to estimate a probability that targeted
fields 106 will be infested. The unsupervised learning engine 144
generates a risk assessment profile 145 based on this estimate. The
profile 145 is applied to generate the pest occurrence prediction
122 as output data 120.
The unsupervised learning engine 144 is configured to generate
dynamic and adaptive risk assessments as a growing season
progresses that are localized to the observation of a pest presence
102 in reported fields 104, as applied to predictions of a pest
presence in nearby targeted fields 106. As additional crowd-sourced
observations containing pest presence data 111 are ingested for
fields 104, the unsupervised learning engine 144 refines its risk
assessment profile 145 by continually modeling the weighted vectors
of field variables that aid in identifying predictors 160 from
descriptors 161. Accordingly, the statistical processes 146 and
other mathematical processes 147 may update the risk assessment
profile 145 by continually and automatically determining the
importance of each predictor 160 from the set of descriptors 161
identified in the expected pest-environment relationship for
modeling the field-specific risk of pest occurrence in the one or
more targeted fields 106.
The data processing components 140 may therefore also be thought of
as performing a customized modeling approach for assessing a
field-specific risk, and for predicting a pest infestation or
presence in a field 106, for a particular time period. The model
100 estimates the probability that targeted fields 106 will become
infested by calculating a similarity of each un-infested, targeted
field 106 to a reporting, or infested, field 104. As noted above,
these calculations can be accomplished in several different ways,
and using several different methods (or, an ensemble of different
methods).
The unsupervised learning engine 144 is configured to construct a
localized and adaptive infestation suitability model 148 based on
the environmental and crop management predictors 160 to map and
determine a risk of pest occurrence in the one or more targeted
fields 106. This infestation suitability model 148 is based on pest
presence data 111 in the reporting fields 104, and may utilize one
or more different methods of mapping infestation suitability from
this pest presence data 111. One such method involves calculating
some measure of similarity in the one or more targeted fields 106
to the infested, reporting fields 104, based on environmental and
crop management predictors 160 that describe past crop/field
management and environmental conditions reported or experience in
the reporting fields 104.
There are several different approaches that may be utilized for
this method of calculating measures of similarity. One such
approach may involve assigning coefficients of multivariate
similarity 149 to the environmental and crop management predictors
160 for the location where the pest 102 has been observed by
assigning importance to predictors 160 that particularly influenced
the pest infestation reported in the pest presence data 111. One
way to develop these coefficients is to examine a relative
likelihood of a pest presence in the one or more targeted field 106
based on a comparison of characteristics of a targeted field 106 to
the characteristics to the group of known infested, reporting
fields 104, with knowledge of the particular pest 102 and other
factors such as weather and pest biology. Another way to develop
these coefficients is to assign weights to vectors that influence
environmental and crop management predictors 160. Another approach
to calculating measures of similarity may involve developing
profile methods that are based on ranges of descriptive
environmental and crop management predictors 160 for the location
where the pest 102 has been observed.
Another method for mapping and determining infestation suitability
from pest presence data 111 involves modeling pest presence versus
availability of habitat, and involves characterizing available
habitat by a sample of other locations (or, alternatively, from a
more complete census). This method may employ many different
statistical processes 146 that apply to analysis of a binary
classification (i.e. presence or absence). Samples of locations
having a reported pest presence provide the most suitable data for
developing an infestation suitability model 148. Nonetheless, the
present invention contemplates that fields without reports of a
pest presence may be used in a reasonable sampling of other
locations, particularly where an agriculture retailer or crop
consultant working on a large number of fields in an area needs
information on all interested fields.
Still another method for mapping and determining infestation
suitability from pest presence data 111 is an "infestation
suitability" model 148 based on expert opinion to assign weights
to, and define, transfer mapping functions of environmental and
crop management characteristics of fields. Transfer mapping
functions describe the relationship between the likelihood of
infestation and the value of an environmental or crop management
predictor 160 (i.e. more rain, then more likely). Expert opinion
may be combined with crowd-sourced information for a more accurate
representation of conditions leading to the pest presence.
Regardless, data collected from the infested, reported fields 104
may be used to modify a priori transfer mapping functions.
An ensemble approach that combines one or more of the methods
described herein may also be used to increase accuracy from the
prediction of pest occurrence 122 generated from the risk
assessment profile 145. Regardless, the present invention may also
develop one or more additional artificial intelligence models 165
that are configured to automatically and continually analyze the
input data 110 to ascertain the environmental and crop management
predictors 160 from the set of descriptors 161 for the unsupervised
learning engine 144, as well as to ascertain which pest biological
characteristics will assist in constructing selected predictors 160
that serve to initialize the infestation suitability model 148 in
the unsupervised learning engine 144.
The present invention also contemplates that the unsupervised
learning engine 144 may include applying one or more methods to
measure the similarity of time series of one or more weather
variables for developing the infestation suitability model 148.
Such methods involve calculating the similarity of time series of
weather data, which may be constructed for each field based on
calendar date or crop date (for example, a number of days since
planting, or days since a certain crop growth stage). Such a
calculation of the similarity of time series of weather day may
also serve as one or more of the environmental and crop management
predictors 160, in addition to a separate step in the overall
unsupervised learning engine 144.
In a further method, the unsupervised learning engine 144 may
calculate a similarity from a percentage of fields that have the
same value as the targeted field 106 according to variables from
one or more weighted vector categories. The unsupervised learning
engine 144 may also calculate a probability density function for
each variable and assign a similarity based on where the value of
the variable falls within the probability density function. In
another example, the model 100 may calculate a single measure of
similarity for each field as weighted calculation from the
similarity measures of individual factors. The vector weights would
indicate the significance of the individual factor in promoting
pest presence. The model 100 may also assign pre-defined risk
category to the similarity measure for each field 106, or rank the
fields 106 by their similarity to the infested field 104.
Other methods of analyzing input data may also be employed by the
crowd-sourced pest and disease model 100. In another example, a
Bayesian approach may be applied to update pre-existing models
developed by the artificial intelligence modeling portion 165 of
the present invention, as more fields 104 are reported as infested.
The present invention may also examine similarities in patterns of
important weather variables leading up to pest presence to an
additional layer of accuracy to the model 100. This incorporates
methods of calculating the similarity of time series data, and adds
a further dimension by enabling a look-back at conditions present
in the reporting field 104 that impacted development of the pest
102. Such a time-series look-back measures similarities in
time-series data sets, and enables pattern-matching of attributes
such as weather over time in the targeted field 106 to the pattern
in fields where the pest 102 has been observed.
One exemplary application of a time-series look-back in the present
invention is as follows. Stewart's disease is a bacterial disease
affecting corn crops that is spread by corn flea beetle. Warm
winter air temperatures in December, January, and February may
increase the survival of corn flea beetle and result in greater
transmission of the bacterium in the following growing season. The
present invention would not need a model of winter survival of corn
flea beetle; instead the model looks for fields that had similar
winter weather, along with recent conditions conducive for disease
infection, to develop a prediction of pest occurrence in the coming
growing season.
In addition to the approaches described above, the present
invention may further incorporate one or more existing modeling
approaches (or, an ensemble of such approaches) that may be
suitable for identifying the set of descriptors 161, and selecting
environmental and crop management predictors 160, that are used to
construct the localized and adaptive infestation suitability model
148 for determining a risk of pest occurrence in the one or more
targeted fields 106. Such approaches include an envelope method,
such as BIOCLIM, and distance-based methods such as DOMAIN and
LIVES that assess possible infestation sites in terms of
environmental similarity to areas with a known pest presence. Other
models may also be utilized, particularly where additional
information such as absence data is incorporated, for example
regression-type models may also be applied, such as multivariate
adaptive regression splines (MARS), regression trees, generalized
additive models (GAMs), generalized dissimilarity models, and
generalized linear models. Other machine learning models may also
be suitable, such as maximum entropy models (MAXENT and MAXENT-T)
and boosted decision/regression trees or stochastic gradient
boosting.
Regardless of the method or approach employed to arrive at selected
predictors 160, the model 100 analyzes the set of descriptors
defining a similarity comparison between the reporting field 104
and the targeted fields 106 from the expected pest-environment
relationship to create the risk assessment profile 145 for the
targeted field 106 from selected environmental and crop management
predictors 160. Risk may be assessed in a number of different ways
using the risk assessment profile 145. For example, a field 106 may
be rated as low risk due to forecasted weather that is not similar
in ways conducive to pest presence. A low risk may attributed where
a resistant crop or seed variety was planted in the field 106. A
low risk may also be assigned where the crop is not predicted by a
crop growth model to be at a susceptible growth stage that matches
with plant biology, as a further example of how pest biology data
may be thought of as a filter for matched interactions between
weather and crop management data. Conversely, a field may be rated
at high risk where the pest is very mobile, and a known infested
field is in close nearby proximity. Risk may further be assessed as
a numerical value, or in one or more indicators specifically
tailored to particular fields, crops, growers, or users. The
present invention may apply the risk assessment profile 145 to
identify suitable windows of opportunity for performing certain
cultivation tasks, or applying treatments, to avoid or mitigate
damage from infestation. Alternatively, users of the model 100 may
also receive, as an output indicator, a measure of likelihood or
risk of a particular pest for the targeted field 106. The model 100
may also generate an indication or reason of why the targeted field
106 is at high risk (such as planting date, weather, pest life
cycle, seed variety, etc.) It is therefore to be understood that
many different methods and approaches of applying the risk profile
to assess risk may be utilized and are within the scope of the
present invention.
Certain types of the input data 110, for example biological
characteristics of a pest 102 in pest and disease data 115, may be
used to filter the weather and crop management information to
determine and select relevant environmental and crop management
predictors 160 impacting a pest-environment relationship 143 for
that field 102, or to inform or assign weights to vectors of the
field variables when such a method is used. One exemplary situation
is where a pest problem in a field is the result of the
intersection of the crop at a vulnerable growth stage and the pest
reaching the damaging growth stage. However, the optimal weather,
management and field conditions that result in crop reaching a
vulnerable growth stage (a particular variety, planting date and
weather conditions interacting with the soil type) are not the same
as the optimal conditions for the pest to reach the damaging growth
stage (presence of a particular type and amount to crop residue for
overwintering of the pest). For example, a large presence of crop
residue may translate to a higher population of the overwintering
stage of a pest, but the amount of crop residue can influence
warming of soil in the spring and consequently the speed of
emergence of the crop. There may be further complications if the
pest has natural enemies and development and abundance of the
natural enemy is driven by different weather conditions than crop
growth or emergence of the overwintering pest from the soil. The
modeling paradigms of the present invention aid uses in such
situations, because the severity of damage from a pest and
consequently the need for control may be influenced in complex ways
by weather that are best captured by similarity analyses, rather
than sorting out causal factors. In one specific example, in common
smut infestations of corn, moisture is needed for fungal spores to
germinate and penetrate the host, so the infested field will have
recent rainfall or high relative humidity. However, the spores of
smut fungus are able to only infect tissue that has been damaged.
Appropriate damage may be caused by blowing soil particles or hail.
If a crowd-sourced observation of corn smut is associated with hail
or high winds in the field, this would be reflected in weather data
for the current growing season. Fields with matching weather would
be predicted to likely be at risk of corn smut infestation.
It is to be understood further that many other types of output data
120 are possible. The risk assessment profile 145, as noted above,
may be used to generate specific predictions 122 of pest
infestation in targeted fields 106, and one or both of these may be
further used to suggest, recommend, or generate one or more
management actions, either before or after infestation, to address
a pest infestation and/or mitigate the impact. Additionally, many
different users and uses of this output data 120 are possible.
Output data 120 may therefore be used to perform several functions,
either directly or through other systems, hardware, software,
devices, services (such as the advisory services 150 described
below), and application programming interfaces 170.
Examples of management actions include notifications to begin
scouting targeted fields 106 to confirm presence of a pest 102,
and/or to confirm that the density or extent warrants a control
action. Scouting may also be advised to confirm presence of a pest
102, and/or for observation and planning for the next growing
season in targeted fields 106. Other actions include notices to set
out traps (for example, for insects), to control the infestation,
such as through a pesticide application, or other practice such as
tillage, or apply other preventative treatment, for example where
the pest 102 is a pathogen. Crop-specific actions may include
delayed planting and use of a resistant seed variety to mitigate
any impact from the pest 102. Notifications may also be provided
directly to farm equipment operating in a targeted field 106. For
example, a notification may be provided directly to tillage
equipment to adjust or stop tilling of a targeted field 106, or to
irrigation equipment operating in a targeted field 106 to adjust a
timing or type of artificial precipitation used, or a direction of
application.
The crowd-sourced pest and disease model 100 may also enable one or
more specific application programming interface (API) modules 170
to provide all of the modular services described above and generate
specific outcomes, such as one or more specific advisory services
150. Alternatively, these services 150 may be provided directly by
the present invention. Regardless, such a service 150 or API 170
may be tailored to provide specific management actions.
For example, the present invention may provide a crop and soil
conditions advisory 151 regarding targeted fields 106 that include
information beyond the notifications described above. Such an
advisory service 151 may model possible crop and soil damage from
infestation of a pest 102 and provide analytics of such damage,
such as for example an economic impact on a crop in the current
growing season, or an impact on soils from treatments applied to
crops to control the pest 102.
The present invention may also provide a contamination advisory
service 152 for crops, soils, and groundwater or aquifers that is
provided to owners of fields, growers of crops, and other
responsible entities in relation to targeted fields 106. Such a
service may advise on tillage practices, for example where the pest
102 is a pathogen such as an avian influenza virus that is able to
survive very cold soil temperatures. Soils in targeted fields 106
that are proximate to fields, reporting or otherwise, along
migratory bird routes may become infected from bird droppings carry
such a pest 102. Tillage of contaminated fields may create airborne
particles that are easily spread to other fields. Such an advisory
152 may model the use of certain field equipment, and/or tillage
timing and conduct, and may be dependent on a variety of factors.
Such an advisory 152 may therefore provide tillage practice
analytics to manage pest containment in targeted fields 106.
Many additional agricultural advisories 150 are contemplated.
Examples of advisory services 150 that include other agricultural
management services are a planting and harvest advisory service
153, and a crop and soil nutrient and biological application
advisory service 154, a disease prediction advisory service 155, an
irrigation advisory service 156, and a herd, feed, and rangeland
management advisory service 157. Additional management services may
include a regulatory advisory service 158.
ClearAg and other alerting is still another service 159
contemplated by the present invention. ClearAg is an application
offering a suite of precision agriculture services, includes alerts
that provide subscribers and other users with actionable
information regarding harvest, planting, irrigation, pest/disease,
nutrient, and other issues attendant to crop, field, and soil
management.
All of these advisories 150 are possible with the output data 140,
based on the input data 110 ingested. For example, a regulatory
advisory service 158 may produce an advisory based on the risk
assessment profile 145 where a recommended management action is
application of a particular chemical or treatment to eradicate a
pest 102. Such an advisory 158 may indicate that a soil will have a
high contamination risk of a substance that requires federal or
state reporting after application. Another example of a regulatory
advisory service 158 is an indicator of predicted environmental
impact from runoff following delivery of a chemical treatment to
soils, where irrigation patterns are a known component of crop
management data.
In a further example, an irrigation advisory service 156 may
consider predictions of pest occurrence 122 to inform growers,
landowners, or other responsible parties of irrigation-related pest
mitigation actions, such as the positioning of flood, drip, and
spray irrigation equipment, the timing of their use, and amounts of
artificial precipitation to be applied. In still a further example,
the herd, feed, and rangeland management advisory service 157 may
provide information for reducing a stocking rate or purchase of
additional feed for existing stock for rangeland management to
mitigate a threat of pest infestation.
It is to be noted that advisory services 150 may be provided as a
specific outcome of the present invention where it is configured to
provide all of the modular services described above in a packaged
format, and the advisory services 150 may also be processed from
output data 140 (either directly, or via the API modules 170). It
is further to be understood that many such advisory services 150
and API modules 170 are possible and are within the scope of the
present invention.
As noted above, in one or more additional or alternative
embodiments of the present invention, the present invention may be
thought of as an artificial intelligence approach 165 that may be
applied to develop relationships between the various types of input
data 110 to select variables and identify a set of descriptors 161,
to generate the expected pest-environment relationship 143. The
unsupervised learning engine 144 is configured to automatically and
continually explore relationships between the input data 110 to
select the environmental and crop management predictors 160 based
on the evaluation of multivariate similarity 149 between variable
values as pest presence data 111 in crowd-sourced information is
received. As more and more input data 110 is accumulated,
information can be sub-sampled, and the crowd-sourced pest and
disease model 100 retrained, to develop a more reliable
understanding of infestation suitability in the model 148. Such
modeling may also implicitly yield information as to the importance
of environmental and crop management predictors 160, and any
multivariate coefficients assigned to them. This may be used to
select which predictors 160 are particularly important or
unimportant in the associated process, and thus help to target ways
of improving the crowd-sourced pest and disease model 100 over
time.
FIG. 2 is a flowchart of steps in a process 200 of performing a
crowd-sourced pest model according to another embodiment of the
present invention. In FIG. 2, the process 200 receives pest
presence data 111 in anonymous crowd-sourced observations for a
reporting field 104 at step 202, and proceeds by correlating
positional coordinates of both reporting fields 104 and targeted
fields 106 at step 204. If there is no correlation between
reporting fields and targeted fields 106, the process 200
terminates; if there is correlation, the process 200 proceeds with
step 206 to identify the pest 102, and collect additional the input
data 110 needed for modeling the risk assessment profile 145 the
crowd-sourced pest and disease model 100. The process 200 may also
include receiving a request to generate a risk assessment profile
145 for a targeted field 106.
The process 200 then proceeds with identifying field variables to
develop the expected relationship between pest and environmental
conditions in reporting fields 104 at step 208, as described in
detail above. The model 100 then begins, at step 210, with
analyzing field variables to compare characteristics of reporting
fields 104 and targeted fields 106 in the unsupervised learning
engine 144. At step 212, the unsupervised learning engine 144
creates a risk assessment profile 145 for the targeted field(s)
106. The process 200 then generates one or both of a probability
and prediction of pest occurrence 122 at step 214. The process 200
may repeat continuously, as additional presence data 111 is
analyzed and ingested in anonymous crowd-sourced observations, so
as to automatically and continually refine the output data 120.
Meteorological and climatological data 114 may be collected from
many different sources of weather information to provide one or
more of the recent and current field-level weather data and
short-term weather forecast data, for example as data that is
complementary to the data assimilation systems and forecasting
systems noted below. As noted above, weather information may be
ingested into the present invention in either raw or processed
form, from many different sources. Such sources of weather
information may include data from both in-situ and remotely-sensed
observation platforms. For example, numerical weather prediction
models (NWP) and/or surface networks may be combined with data from
weather radars and satellites to reconstruct the current and
near-term forecasted weather conditions on any particular area to
be analyzed. There are numerous industry NWP models available, and
any such models may be used as sources of meteorological data in
the present invention. Examples of NWP models at least include RUC
(Rapid Update Cycle), WRF (Weather Research and Forecasting Model),
GFS (Global Forecast System) (as noted above), and GEM (Global
Environmental Model). Meteorological data is received in real-time,
and may come from several different NWP sources, such as from the
European Centre for Medium-Range Weather Forecasting (ECMWF),
Meteorological Services of Canada's (MSC) Canadian Meteorological
Centre (CMC), as well as the National Oceanic and Atmospheric
Administration's (NOAA) Environmental Modeling Center (EMC), and
many others. Additionally, internally or privately-generated
"mesoscale" NWP models developed from data collected from real-time
feeds to global and localized observation resources may also be
utilized. Such mesoscale numerical weather prediction models may be
specialized in forecasting weather with more local detail than the
models operated at government centers, and therefore contain
smaller-scale data collections than other NWP models used. These
mesoscale models are very useful in characterizing how weather
conditions may vary over small distances and over small increments
of time. The present invention may be configured to ingest data
from all types of NWP models, regardless of whether publicly,
privately, or internally provided or developed.
In one embodiment of the present invention, the model 100 is
applicable within an organizational crowdsourcing environment. An
organization may be, for example, an agriculture retailer, crop
consultant, or large agribusiness enterprise, and therefore the
crowd-sourced observations of presence data 111 may include
intra-organizational information about soils, crop, planting date
and management that have been previously collected across the
multiple fields of the entire organization. An example of an
intra-organizational use in such an embodiment includes multiple
fields in proximity to each other that are owned, operated or
maintained by a common enterprise, and on which different seeds,
crops, and tillage practices are being applied. The one or more
targeted fields 106 may therefore comprise an entire area of
arbitrary size, and it is to be understood that targeted fields 106
may be not contiguous, regardless of the embodiment. Regardless,
the approach used in the present invention to perform the
infestation suitability model 148, such as the environmental and
crop management predictors 160, may rank multiple fields of a
single customer according to likelihood of pest occurrence. Such a
ranking may assist users to prioritize which fields to manage or
visit among a set of fields.
The systems and methods of the present invention may be implemented
in many different computing environments 130. For example, they may
be implemented in conjunction with a special purpose computer, a
programmed microprocessor or microcontroller and peripheral
integrated circuit element(s), an ASIC or other integrated circuit,
a digital signal processor, electronic or logic circuitry such as
discrete element circuit, a programmable logic device or gate array
such as a PLD, PLA, FPGA, PAL, and any comparable means. In
general, any means of implementing the methodology illustrated
herein can be used to implement the various aspects of the present
invention. Exemplary hardware that can be used for the present
invention includes computers, handheld devices, telephones (e.g.,
cellular, Internet enabled, digital, analog, hybrids, and others),
and other such hardware. Some of these devices include processors
(e.g., a single or multiple microprocessors), memory, nonvolatile
storage, input devices, and output devices. Furthermore,
alternative software implementations including, but not limited to,
distributed processing, parallel processing, or virtual machine
processing can also be configured to perform the methods described
herein.
The systems and methods of the present invention may also be
partially implemented in software that can be stored on a storage
medium, executed on programmed general-purpose computer with the
cooperation of a controller and memory, a special purpose computer,
a microprocessor, or the like. In these instances, the systems and
methods of this invention can be implemented as a program embedded
on a mobile device or personal computer through such mediums as an
applet, JAVA.RTM. or CGI script, as a resource residing on one or
more servers or computer workstations, as a routine embedded in a
dedicated measurement system, system component, or the like. The
system can also be implemented by physically incorporating the
system and/or method into a software and/or hardware system.
Additionally, the data processing functions disclosed herein may be
performed by one or more program instructions stored in or executed
by such memory, and further may be performed by one or more modules
configured to carry out those program instructions. Modules are
intended to refer to any known or later developed hardware,
software, firmware, artificial intelligence, fuzzy logic, expert
system or combination of hardware and software that is capable of
performing the data processing functionality described herein.
The foregoing descriptions of embodiments of the present invention
have been presented for the purposes of illustration and
description. It is not intended to be exhaustive or to limit the
invention to the precise forms disclosed. Accordingly, many
alterations, modifications and variations are possible in light of
the above teachings, may be made by those having ordinary skill in
the art without departing from the spirit and scope of the
invention. It is therefore intended that the scope of the invention
be limited not by this detailed description. For example,
notwithstanding the fact that the elements of a claim are set forth
below in a certain combination, it must be expressly understood
that the invention includes other combinations of fewer, more or
different elements, which are disclosed in above even when not
initially claimed in such combinations.
The words used in this specification to describe the invention and
its various embodiments are to be understood not only in the sense
of their commonly defined meanings, but to include by special
definition in this specification structure, material or acts beyond
the scope of the commonly defined meanings. Thus if an element can
be understood in the context of this specification as including
more than one meaning, then its use in a claim must be understood
as being generic to all possible meanings supported by the
specification and by the word itself.
The definitions of the words or elements of the following claims
are, therefore, defined in this specification to include not only
the combination of elements which are literally set forth, but all
equivalent structure, material or acts for performing substantially
the same function in substantially the same way to obtain
substantially the same result. In this sense it is therefore
contemplated that an equivalent substitution of two or more
elements may be made for any one of the elements in the claims
below or that a single element may be substituted for two or more
elements in a claim. Although elements may be described above as
acting in certain combinations and even initially claimed as such,
it is to be expressly understood that one or more elements from a
claimed combination can in some cases be excised from the
combination and that the claimed combination may be directed to a
sub-combination or variation of a sub-combination.
Insubstantial changes from the claimed subject matter as viewed by
a person with ordinary skill in the art, now known or later
devised, are expressly contemplated as being equivalently within
the scope of the claims. Therefore, obvious substitutions now or
later known to one with ordinary skill in the art are defined to be
within the scope of the defined elements.
The claims are thus to be understood to include what is
specifically illustrated and described above, what is conceptually
equivalent, what can be obviously substituted and also what
essentially incorporates the essential idea of the invention.
* * * * *